The information storage potential of DNA and RNA is limited by the selective hydrogen bonding (H-bonding) of the natural base pairs. We aim to circumvent this restriction by designing unnatural base pairs and evolving polymerases that recognize them. We have been exploring base pairs whose interactions are driven by intermolecular forces other than H-bonding, including hydrophobicity, whose contribution to protein structure has been appreciated for decades. During the previous funding period, we successfully completed our specific aims to 1) synthesize and characterize unnatural base pairs; and 2) establish a selection system to evolve DNA polymerases that accept the unnatural substrates. From this work, several replication systems have been developed, including one based on 'self-pairs' formed between two 3-fluorobenzene nucleosides (3FB). The 3FB self-pair is the first unnatural base pair that is processively synthesized with selectivity against all possible mispairs. Preliminary structural characterization implies that the self-pairing is mediated by dipole-dipole forces and an interesting interbase fluorine-hydrogen bond. We have also developed an activity based selection system and in three separate applications shown it to be capable of evolving DNA polymerases that synthesize unnatural DNA. We now propose to build on these successes by first characterizing the structure and replication of 3FB self-pairs and mispairs, and using this data to aid in the design of a better self-pair. We will also evolve DNA polymerases specifically tailored to replicate DNA containing 3FB self-pairs. Finally, we will begin to explore the transcription of the self-pair into unnatural RNA, the next step toward expanding an organism's genetic code, laying the foundation for a semi-synthetic organism. [unreadable] [unreadable]